DISTRIBUTION IN THK ATMOSPHERE OVEK ENGLAND. 265 



Messrs. <!tn.n :m<l 1 1 AHWOOD give a very similar range and phase, but their mean 

 height is only 10'G for the whole set and 10'8 for the British Isles. Dr. WAQNEK 

 gives a somewhat larger range with a minimum in March and a mean height of 10'5. 

 M. RYKATCHEW gives for Pawlowsk about the same range, but his mean height is 

 only 9 '6 km. 



The lower height at Pawlowsk may be explained by the higher latitude. The values 

 at Pawlowsk are included in Messrs. GOLD ;mil HAKWOOD'S and in Dr. WACJNER'S 

 results and reduce their mean, but on the whole tin- < 'ontinental stations are not as 

 far north as England. It does not seem possible that there can be any large 

 systematic error in the calibration of the instruments, 0'5 of a kilometre is at this 

 height equivalent to 14 mm. of pressure, so that it appears that the mean value of 

 H e is greater for England than for the Continent, and this confirms Dr. WAGNER'S 

 statement. He says the H c decreases from south to north, and from the ocean to the 

 Continent. 



Dr. SOHMAUSS pointed out the annual variation in 1909. Whatever the explana- 

 tion of the isothermal region may be, it seems reasonable to expect it to occur always 

 at the same pressure rather than always at the same height. In August the mean 

 temperature of the air, up to the isothermal, is 13 C. above its value in February, 

 and in consequence at 11 km. the same pressure is met with some 0'5 km. higher, 

 but this does not suffice to explain the whole variation. 



The Value of If e for Various Heights of the Barometer. 



Height of barometer ... 738 744 755 7GO 76G 77:3 mm. 



Value of H c 8'4 9'1 10'3 ll'O 12 '4 12 '3 km. 



Number of cases G 12 15 30 41 20 



These figures have been corrected for the annual variation. 



It will be seen that, if we exclude the last group, the relation between H c and the 

 height of the barometer may be taken as practically a linear one. I am inclined to 

 think that, notwithstanding the large number of observations in the 766 mm. group, 

 the value 12 '4 km. for this group is accidentally high. In the lower groups the 

 individual values are very consistent. In the 18 ascents made with the barometer 

 In-low 750 mm. in one only was a value of H,, above 10 km. met with, and the lowest 

 was 7 '5. In the higher groups there is great irregularity, so that it may be said 

 with fair certainty that a low value of He will be found with a low barometer, but 

 the converse does not hold. Of all the quantities relating to the upper air this value 

 of He is the most capricious and uncertain, save during pronounced cyclones. There 

 is evidence that, in the 40 miles between Pyrton Hill and Ditcham Park, it may 

 <lin"t>r by at least 1 km., and that the same change may occur in the same locality in 

 less than an hour. 



1 1' we combine the last two groups we get the rule that a rise of 3 mm. in height of 



VOL. ccxi. A. 2 M 



